In data systems, an application programming interface (“APIs”) may be defined and created in which data write requests are generated and sent in synchronization with received data store commands, such that each data write request is conveyed and stored in the system as a result of a respective data store command. This approach permits synchronization and low-latency read-back of the data, but also increases communication load in the system. Conversely, an API may be defined and created in which two or more data write requests are aggregated and conveyed together for storage in the system. This approach reduces communication load, but can adversely affect latency if there is a read request for data previously included in a store command but which has not been actually stored yet due to aggregation of write requests. Needed are systems and methods that can reduce communication load without significantly increasing system latency.